This folder contains the scripts used for comparing prediction accuracies between matched AA and WA.

## Calculate accuracy metrics for matched AA and WA separately

Related function: `HCP_KRR_acc_AAvsWA_matchedBehavior.m`. Example (matlab):

```matlab
HCP_KRR_acc_AAvsWA_matchedBehavior( '/your/kernel/regression/output/dir/', 400, 10, ...
   '/your/AA/splits/dir/', '/your/WA/splits/dir/', '/your/selected/seed_behavior_comb/lists/dir/', ...
   '/output/predictive_COD.mat', '/output/Pearson_correlation.mat', ...
   '/list/of/behaviors/with/enough/matched_AAandWA/txt', '/full/behavior/list.txt', ...
   '/full/subject/list.txt' )
```

## Statistical test for the accuracy difference between AA and WA

Related function: `HCP_PermTest_AAvsWA.m`. Example (matlab):

```matlab
HCP_PermTest_AAvsWA('/AAvsWA/accuracy.mat', '/focused/behavioral/list.txt', ...
   'predictive_COD', '/output/statistical/test.mat')
```